The present study investigates the calcium-independent component of the slow AHP in lamprey spinal neurons, its underlying mechanisms and functional significance. This non-KCa-sAHP is readily discernable even after a single action potential (see Cangiano et al. 2002), and it will furthermore sum very effectively during repetitive firing and even more so than the KCa component, so that after a spike train it may constitute more than half of the summed control sAHP (Figs 1 and 2). Most likely the non-KCa-sAHP will therefore play a significant role for firing properties during burst firing.
The non-calcium-dependent sAHP is due to a sodium-activated potassium current (KNa), mediated by the Slack gene
The nature of the calcium-independent component of the sAHP was investigated in several steps. We first explored the possibility that the non-KCa-sAHP could be due to a sodium-dependent potassium current (KNa). We therefore analysed the effect of replacing sodium with lithium ions. Lithium ions flow through voltage-gated sodium channels almost as easily as sodium ions (Hille, 1972), and action potentials can be evoked in lithium without any marked change in threshold (see Bischoff et al. 1998). Lithium has, however, been reported not to activate KNa currents (Safronov & Vogel, 1996; Bischoff et al. 1998). Accordingly, when sodium ions were replaced with lithium ions in the present study, the non-KCa-sAHP was markedly reduced in amplitude (Fig. 3). Thus, the current underlying the non-KCa-sAHP appears to be sodium dependent. Next, we could conclude that this current is most likely a potassium current (Fig. 3). The responsible potassium current was further investigated in experiments with blockers (TEA, catechol) of voltage-dependent potassium channels (Fig. 4), which, rather than blocking the sAHP, caused a marked amplitude increase. This increase was proportional to the enhanced action potential area due to spike broadening, with a concomitant increase in Na+ entry, occurring with these blockers. The underlying potassium current is thus different from the delayed rectifier current or the transient potassium current, which are blocked by TEA and catechol, respectively.
The remote possibility that a specific activation of the sodium–potassium pump could underlie the non-KCa-sAHP was also investigated (Fig. 5; see Thompson & Prince, 1986; Morita et al. 1993; Scuri et al. 2002). The reduction of the sAHP seen with ouabain was always accompanied by a change in reversal potential for the sAHP and in resting membrane potential, and can thus be accounted for by the change of the potassium reversal potential. Furthermore, a maintained sAHP could be readily revealed during ouabain treatment by depolarizing the cell above the reversal potential. Thus, a specific activation of the sodium–potassium pump appears not to be responsible for generating the non-KCa-sAHP in this case, and possibly not in other studies that have observed a change in AHP amplitude when applying ouabain (see Parker et al. 1996).
All evidence thus indicates that the calcium-independent component of the sAHP is due to a sodium-dependent potassium current, a KNa current. The involvement of a KNa current was further corroborated in experiments using quinidine, which is reported to effectively block both cloned subtypes of KNa channels (Slick (Slo2.1) and Slack (Slo2.2)) (Bhattacharjee et al. 2003).
The immunohistochemical data (Fig. 6) provide support for the presence of a Slack-like subtype of KNa channel in the lamprey spinal cord, demonstrating distinct Slack immunoreactivity in the soma and proximal dendrites of neurons that also showed the KNa-sAHP. Immunoblotting confirmed the presence of the Slack protein in the lamprey CNS. It may further be noted that the peptide region of Slack chosen to generate the antibody used here is highly conserved among multiple species (i.e. the sequence is identical in the rat, mouse, human, cow, elephant, dog, macaque and chimpanzee orthologues of the Slack channel). We also utilized the fact that the sea lamprey (Petromyzon marinus) genome at the moment is being sequenced, although not yet assembled. Using the sequence similarity program Discontiguous Megablast (http://www.ncbi.nlm.nih.gov/blast) and the rat Slack mRNA sequence (accession number AY884213) on the Whole Genome Shotgun (WGS) sequences from Petromyzon, we were able to pick up 18 of 32 exons (data not shown). This adds further support for the presence of a Slack-like KNa channel also in lamprey.
Additional evidence for an involvement of a Slack-like subtype of KNa channel is provided by the experiment with Cl− injection. Slick is activated by both Na+ and Cl−, and is even more sensitive to Cl− than to Na+, while the Slack subtype (Yuan et al. 2003) shows much less Cl− sensitivity and higher Na+ sensitivity (Bhattacharjee et al. 2003). The KNa-sAHP was not affected by Cl− injection and, with these data taken together, the Slack subtype therefore appears as the most likely candidate underlying the KNa-sAHP. To our knowledge, this is the first demonstration of a functional role for the Slack gene in contributing to the sAHP.
The involvement of KNa currents in the generation of different types of slow afterhyperpolarizations has been described in a few other systems. A sodium-dependent sAHP seen after spike train stimulation has been reported in rat motoneurons (Safronov & Vogel, 1996), and in intrinsically bursting neurons of rat neocortex (Franceschetti et al. 2003). In perigeniculate neurons (ferret), spindle wave activity is followed by a long-lasting AHP, the slowest component of which (several seconds) has been suggested to rely on a sodium-dependent K+ current (Kim & McCormick, 1998). The fact that in the present study the sodium entry during a single action potential is sufficient to activate the KNa channels stands in contrast to previous reports (Dryer, 1994; Safronov & Vogel, 1996; Franceschetti et al. 2003), possibly indicating close proximity between the source of sodium entry and the KNa channels and/or a high sodium-affinity of the channel (see Koh et al. 1994; Dryer, 2003).
The fast, transient KNa current occurring during the action potential (Hess et al. 2007) may be due to a rapid activation of KNa channels in immediate proximity to the point of sodium entry, while the KNa-sAHP could be generated by activation of KNa channels located somewhat further away, but still close enough to allow activation upon a single spike. The difference in time course between the fast KNa current during the action potential and the slower KNa current underlying the KNa-sAHP might then reflect the dynamics of diffusion of sodium ions along the membrane. Interestingly, the immunohistochemical analysis revealed expression of the Slack protein in both soma and dendrites of lamprey spinal neurons.
The KNa-sAHP and spike frequency regulation
In the absence of the KCa-component of the sAHP, as when blocked by a modulator, the steady state frequency was markedly increased, while spike frequency adaptation was still present, albeit less pronounced (Fig. 7). Thus, without the KCa-sAHP the cell will fire at a higher rate in response to the same excitatory input, but at the same time the presence of the KNa-sAHP maintains the ability for a limited spike frequency adaptation. The KNa-sAHP may therefore be viewed as a safety mechanism; in the event of a down-regulation of the KCa-component due to modulator action, the neuron will still be capable of regulating its firing behaviour.
Summation of the sAHP during rhythmic burst firing
The sAHP, with its major influence on neuronal firing properties, also plays a key role in the operation of a burst-generating neuronal network like the lamprey locomotor CPG (see Grillner et al. 2001; Grillner, 2003). To explore the summation effects of the sAHP during burst firing, this was simulated by repetitive stimulation with pulse trains (Fig. 8). The control sAHP recovered back to baseline between trains at the low burst rate tested (2 Hz), but at higher rates the prominent summation of the sAHP resulted in a net hyperpolarization during the period of stimulation and a slow recovery after its termination. With only the KNa-component of the sAHP operating, summation became significant at higher burst rates (8 Hz), and the net hyperpolarization was less pronounced. The KNa-sAHP might thus be of particular significance during repetitive firing at higher frequencies, as will occur during faster rates of locomotor activity, even though the neuron may fire only a few spikes during the short burst.
Differential modulation of the sAHP
With the sAHP in lamprey spinal neurons being the target for several different modulatory systems, intrinsic to the spinal cord (5-HT/dopamine, GABAB), it was of interest to investigate whether the action of these systems is restricted to an influence on the KCa component via their blocking effect on calcium entry (Matsushima et al. 1993; Schotland et al. 1995; Hill et al. 2003), or if the KNa component is also subject to modulation. Neither 5-HT, nor dopamine nor GABAB receptor activation resulted in any reduction of the KNa-sAHP (Fig. 9), nor did metabotropic glutamate receptor activation. The KNa-component therefore may constitute a robust baseline mechanism, providing the neuron with basic membrane properties for the regulation of firing behaviour, and which is not affected by these modulators that play an important role in the spinal networks.
Implications for the operation of the spinal locomotor CPG
Several modulatory systems of the lamprey spinal cord (5-HT/dopamine, GABA, mGluR) are active during locomotor network activity, and they can thus be regarded as an integral part of the network (see Grillner, 2003). This implies that, for instance, the 5-HT system would act to reduce the KCa-component of the sAHP during network activity, causing less AHP summation and prolonged bursts and thereby contributing to the maintenance of a slow, regular burst rhythm (see Wallén et al. 1989a,b; El Manira et al. 1994). During stronger network drive and faster rhythmic activity, with shorter bursts of higher spike frequency, the KNa-sAHP may become more important and even dominant, particularly if the KCa component is reduced by modulatory action. The KNa-sAHP could then again be viewed as a potential safety mechanism, assuring adequate frequency regulation also when the KCa component is small.